1. We should probably check for "Oops" messages in dmesg just as we do for "BUG" 2. We might want to check for Oops/BUG since boot before the test and not run if the machine's kernel might already be unstable